Nextcloud AIO containers don't start containers

Hi,

I have nextcloud aio installed on unraid, from community apps.
But my master container won’t start my nextcloud containers.

Of course, i can select autostart on unraid, but then the redis container get an error, because it needs to be started from master container.

:signal-handler (1709895857) Received SIGTERM scheduling shutdown…
8:M 08 Mar 2024 13:04:17.361 # Redis is now ready to exit, bye bye…

Hi, can you post the mastercontainer logs here?

1 Like

Sure.

https://pastebin.com/raw/sVzgSPCw

Can you click on start containers and if it should fail post the mastercontainer logs again?

From the nextcloud master container, if I press start, the containers will start, with no errors on redis container.

But the master container is set as autostart in unraid. shouldn’t the master container automatically start the rest of the AIO containers?

If I select all of them to autostart in unraid, then an error appears in the redis container.

Anyway, here are the requested logs: https://pastebin.com/raw/GSNziXj0

Hey @szaimen are there any new informations here? I am having the exact same issue. My master container doesnt start the other containers. Shouldnt it do that?

Hi, what exactly is your issue? Can you post the mastercontainer logs here?

1 Like

Sorry for the late reply. The issue is, that the master container is set to autostart in unraid, but the master container doesnt start all the other containers. I have to manually access the aio-webui and start all containers from there.

1 Like

Hey @szaimen is this intended behaviour? Is there anything i can change, to make the container autostart?` the others?

The child containers should automatically be started together with the docker daemon. The RestartPolicy for them is set to unless-stopped.

Maybe you can describe more in detail what exactly you are doing and what it is happening compared to what you are expecting.

I tihnk t hat works a bit different for Unraid.
For Unraid you can configure Autostart (which more or less is the “unless-stopped” option here:


with the toggle on the right.

Here only the mastercontainer is set to autostart. The others arent. Also i cant just set them to autostart, as it seems like, they have a specific order in which they autostart, which i havent figured out yet. It would be easier, if there was an option for the mastercontainer to start the child-containers with it. The master container has ro rights on the docker socket, so should be doable.

Honestly, I think unraid should rather rely on the standard feature of docker namely RestartPolicy instead of implementing a custom solution.

1 Like

Maybe, but it looks like it doesnt. Therefore there needs to be another solution

Wouldnt it be possible to have an environment variable to configure, so the mastercontainer autostarts (or maybe updates and autostarts) all the child containers, just in case someone is not able to easily configure the autostart like on unraid, there still would be the variable

No, but you might be able create a script that does this like so: GitHub - nextcloud/all-in-one: 📦 The official Nextcloud installation method. Provides easy deployment and maintenance with most features included in this one Nextcloud instance.

1 Like

In my opinion, there should be an official simple way to do this. I know alot of people who use unraid with nextcloud-aio and they all face this problem.

The Unraid docs suggest you can specify the order/etc: Docker Management | Unraid Docs

Is that not the case?

I am wondering how is unraid storing this config flag to autostart the container? Is it stored in a custom database inside the docker management tool or somewhere else like e.g. as a label for the container? Without this information we cannot do anything in AIO.

The autostart order is in a file called “unraid-autostart” in the docker data directory (or image) thats configured through unraids docker settings

I fear there is no way then for AIO to automatically configure this for you and also no manual way for AIO since unraids implementation is done in this way. So you need to enable this flag manually via this unraid interface for all containers.